WHAT’S BEING CLAIMED:
- The Pentagon’s high tech research agency, DARPA, invests $2 billion into Artificial Intelligence (AI) system research.
- DARPA laid the foundations for the US Defense Department for cutting-edge weaponry to be at par with Russian and Chinese military forces.
- The agency aims to provide advanced solutions to military problems by integrating AI into US weaponry.
The Pentagon’s high tech research arm, The Defense Advanced Research Projects Agency (DARPA), plans to focus its efforts on artificial intelligence. The announcement was made Friday during its 60th-anniversary conference, revealing its $2 billion investment in the field.
Artificial intelligence has been on the up and up recently, especially in technology and business trends. AI works to enable machines to perform tasks that are usually accomplished by humans. One example is when Google showed observers an AI system capable of sounding completely human while calling a restaurant and booking a reservation.
Now, machines have the increasing capability to process and understand speech, words, and images. Familiar products with this kind of technology are Amazon’s Alexa, Apple’s Siri, and self-driving vans operated by Waymo.
John Everett, deputy director of DARPA’s Information Innovation office believes that they can “accelerate two decades of progress into five years.”
A lot of the country’s biggest companies rely on AI technology to stay ahead of their competitors. Different national governments like Canada, China, India, and France see AI as an essential part of growing their economies, especially in the 21st century. Hence, they want to prioritize working on AI.
DARPA aims to use their investment to create systems that possess common sense, contextual awareness, and efficiency when it comes to energy use.
However, the AI industry also has a problem when it comes to hype. Huge data sets for algorithms are what machines rely on to learn. It can take millions of pictures of cats before a computer will be able to recognize a single cat from another photo.
AI systems like these often need countless computer chips processing data for long periods of time, weeks even, before machine-learning can occur.
According to Everett, “Machine learning is remarkably inefficient. It can do amazing things, but it’s also remarkable what it can’t do.”
Humans, as compared to machines, can learn something from observing a single example. DARPA wants to embrace AI systems that emulate methods of human learning.
“If we get positive results and they’re important, and they’re relevant to the military and national security, we’re not going to stop,” Everett said. Ultimately, DARPA may put even more money into developing AI systems and technology.